WECC JSIS Meeting€¦ · Post Processing for list of bad measurement Substation Mod el (CIM) Data Record ing Sync h rop as Task 3.2 Pseud o-Synchrophasor Task 3.1 Alarming to Id

|

WECC JSIS Meeting

Heng (Kevin) Chen and Lin Zhang, EPG

Yanfeng Gong and Qiushi Wang, AEP

May 17, 2018

|

▪ Acknowledgment: This material is based upon work supported by the Department ofEnergy under Award Number DE-OE0000850.

▪ Disclaimer: This report was prepared as an account of work sponsored by an agencyof the United States Government. Neither the United States Government nor anyagency thereof, nor any of their employees, makes any warranty, express or implied,or assumes any legal liability or responsibility for the accuracy, completeness, orusefulness of any information, apparatus, product, or process disclosed, orrepresents that its use would not infringe privately owned rights. Reference hereinto any specific commercial product, process, or service by trade name, trademark,manufacturer, or otherwise does not necessarily constitute or imply itsendorsement, recommendation, or favoring by the United States Government or anyagency thereof. The views and opinions of authors expressed herein do notnecessarily state or reflect those of the United States Government or any agencythereof.

© Electric Power Group 2018. All rights reserved1

| 2

▪ Project Introduction

▪ Background

▪ Technical Merit

▪ Technical Approach

▪ Research and System Design

▪ Testing

– Preliminary Simulation Studies

– AEP PMU Deployment and PSCAD Simulation Studies

– 1 Hour Field PMU Data

▪ Current Status & Next Steps

▪ Q&A

© Electric Power Group 2018. All rights reserved

|

▪ DOE/OE and DOE/NETL

– Phil Overholt, Program Manager and Alicia Dalton-Tingler, Project Officer

▪ American Electric Power (AEP) – Sub-recipient

– Project Manager / Alternate – Carlos Casablanca / Yanfeng Gong

▪ Professor Anjan Bose (Washington State University)

– Technical Advisor

▪ Electric Power Group, LLC

– Principal Investigators – Kevin Chen, Lin Zhang

– Key Project Personnel – Ken Martin, Simon Mo, Tingyang Zhang, Neeraj Nayak, Joshua Chynoweth

© Electric Power Group 2018. All rights reserved 3

|

▪ Billions of dollars on transmission and distribution assets

▪ Key substation assets include transformers, circuit breakers,instrument transformers (CTs, PTs, CCVTs) and IntelligentElectronic Device (Relays, PMU, DFRs)

▪ Synchrophasor measurement systems have been widelyinstalled in the North American power grids over the lastdecade

▪ Data from such assets can be used for asset health monitoringand take proactive steps to prevent equipment failure

▪ Proper functioning of substation assets is critical for powersystem operations, reliability and personnel safety


|

▪ Monitor the status and health of substation equipment

▪ Provide early warning indications for potential malfunctioning equipment

▪ Proactively replacement and repair before equipment is damaged

▪ Reduce utility’s forced outage of equipment

▪ Reduce utility’s operating and maintenance costs

Example of failing CCVT in a substation

Example of CCVT voltage signals at Dominion


|

Data from substation will be provided by utility partners▪

Leverage existing synchrophasor technology▪

Research new algorithms in this project▪

Validate at cost share partner substation locations▪

Adapt for general commercial use at other utilities▪

Central Processing: Data sent from substations to central site▪


Provided by Utility Host Provided by EPG

Individual CT/PT timestamped sample

value or phasor

Three PhaseVoltage and Current

Waveform (1, 2, …, m)

Task 3.3

Relay/PMUPost Processing for list of bad measurement

Substation Model (CIM)

Data Recording

SynchrophasorTask 3.2

Pseudo-Synchrophasor

Task 3.1

Alarming to Identify failing Equipment

Visualization DisplaySLSE Engine

DataNXT

Utility Customized One-line Diagram

Task 3.2

Signal Mapping to Substation Model Substation

Network Model Integration

Synchrophasor Data Gateway

One-Line Diagram Kit

EPG component

To be developed

From utility host

Customized One-Line Diagram Integration

Phasor Converter

|

▪ Local Processing at substations: Results sent to asset monitoring center


Provided by Utility Host Provided by EPG

Physical Box In Substation – Hardened PC

Software in Data Center

Individual CT/PT timestamped sample

value or phasor

Three PhaseVoltage and Current

Waveform (1, 2, …, m)

Phasor Convertor

Relay/PMU

Substation Model (CIM)

Data Recording

Synchrophasor

Visualization DisplaySLSE Engine

Substation DataNXT

Utility Customized One-line Diagram

Signal Mapping to Substation Model Substation

Network Model Integration

Synchrophasor Data Gateway


EPG component

To be developed

From utility host

Customized One-Line Diagram Integration

Data Center DataNXT

Post Processing for list of bad measurement


Post Processing for list of bad measurement


|

# SOPO Tasks and Subtasks Planned Timeline

1.0 Project Management and Planning March – April 2017

2.0 Planning, Research, Design, and Specification April 2017 – March 2019

2.1 Overall Project Management April 2017 – March 2019

2.2 Research and Scoping Study March – June 2017

2.3 Functional Design and Design Specifications March – July 2017

3.0 Development, Testing, and Demonstration July 2017 – August 2018

3.1 Pseudo-Synchrophasor Data July – December 2017

3.2 Field Synchrophasor Data December 2017 – March 2018

3.3 Sampled Data from Instrument Transformers April – August 2018

4.0 Deployment and Demonstration at Host Utility September 2018 – March 2019

4.1 Product Documentation September – October 2018

4.2 Installation and Integration at Host Utility October – November 2018

4.3 Site Acceptance Testing November – December 2018

4.4 Demonstration at Host Utility January – February 2019

4.5 Training February – March 2019

5.0 Marketing and Outreach September 2018 – March 2019

5.1 Market Research September – December 2018

5.2 Commercialization Plan January – March 2019

5.3 Outreach September 2018 – March 2019

8

* Requested 1 year project extension according to new PMU deployment schedule at 765kV Vassell substation


|© Electric Power Group 2018. All rights reserved9

|© Electric Power Group 2018. All rights reserved 10

Cause of Failure / Failure Modes

CT PT CVT/CCVT

•Loose Connections or Corroded Connections•Shorting of Winding Turns•Turns to Ground Shorting•Open CT secondary•Insulation

•Erosion of insulation, Insulation Failure•Voids in Insulation –Increased moisture content, Partial Discharge – increased dielectric losses•Aging of CT and wiring insulation, Oil Leaks•High Insulation power factor of internal insulation

•Magnetic core saturation

•Ferroresonance•Switching Transients•PT Saturation

•Insulation Failure•High Stress Voltage Difference across some of the windings•Shorting of Adjacent Windings due to insulation failure•Deterioration of Insulations

•Transient Overvoltage's & Lightning surges•Loose Connections

•Failure of one or more capacitor elements in HV stack – Overvoltage and Stress on each capacitor•Failure of one or more capacitor elements in LV grounding stack – decrease in secondary voltage•Failure of intermediate voltage transformer or series reactor – change in phase angle and/or voltage•Failure of Ferroresonance suppression circuit –waveform distortion, changes in phase angle and/or voltage•Multiple element failure can cause explosion –Staff Safety Issues•Failure of filter circuit or spark gaps used for harmonics & transient voltage reduction –causes increased stress on components•External Flashover, failure of other components – expansion membrane, gasket seal•Low oil conditioned due to oil leak – capacitor failure

|

Raw PMU Data•

LSE Data•

Redundant PMUs•

Other Phases•

DFR Data*•

Available Input - Data

Minimal false positive•

Minimal false negative•

Maximize prediction time•

Within Computing •

Constraints

Desired Output – Flag Asset Fail



Data-Driven Statistical

Detection Flag

SLSE Flag

Cross-check AlarmYes

Data-Driven Statistical

Detection Flag

SLSE Flag

Timer Threshold 3

Timer Threshold 1

No

Alert Alert

Timer Threshold 2

Alarm

Timer Threshold 4

Alarm

SLSE and data• -driven statistical

detection flags are cross checked

for consistence

Two different user• -defined timers

are used to track these flags


DataNXT

Asset Monitoring

ServiceSLSE Service


Grafana

InfluxDB

Measurements


|

▪ A typical breaker-and-a-half schema 500 kV substation configuration:– Full observability of current injection and flow

– Breaker currents are as measurement inputs, as well as bus and line voltages


4730

900

902

4598

4724

4588

4586

4731

4728

47324594

V1, I1

V2, I2

V3, I3V4, I4

V5, I5V6, I6

PMU Current Measurement

Breaker Open

Breaker Closed

911910

923

916

919901

|

▪ A 1% noise to the original signal is added as anomaly to the raw measurement V2 voltage magnitude:


Comparison of raw and estimated VM for V2

Substation equipment status alarm:

The alarm points to the PT feeding the voltage signal.

|

▪ Breaker 4598, 4588 and 4728 are open.▪ A 30 degree offset to the original signal is added as anomaly to the raw

measurement V4 voltage angle:

17

Comparison of raw and estimated voltage angles for V4:

Raw voltage magnitude data and its derivative over 5 time increments:



|

▪ 3 new PMUs deployed at West Campus

▪ 3 PMUs planned at Vassell by Sept 2018

▪ Mainly to get breaker current signals


|

Vassell Set • 1: CCVT 8 ScenariosWest Campus Set • 1: CCVT 10 Scenarios


|

• Vassell Set 2: CT 20 Scenarios• West Campus Set 2: CT 22 Scenarios


|

22

▪ The variance is calculated one phase at a time with 3 moving windows

▪ Main window

▪ Delayed Window

▪ Variance Window – Centered data

▪ Square the centered data

▪ Moving average of Squared data

▪ Moving threshold is obtained based on a scaling factor


|

23


|

1C - CCVT 711 - 1 capacitor fails first at 5 s, 2nd capacitor fails after 30sec, in phase A


|25 25

4A - Normal Operation, one CT turn-to-turn shortage occurs at 10sec in phase A


|26 26

11B - A single phase-to-ground bus fault on bus 1 phase A at 10sec, fault duration is 0.06 s, open D1, C1, B1 at t = 10.05s, reclose at t=10.55s.


|

Control chart is a graph or chart with limit lines. There are basically three kinds of

control lines:

• the upper control limit (UCL),

• the central line, and

• the lower control limit (LCL).

The UCL and LCL are calculated based on a 20σ

1. Identifying the maximum and minimum values in 1-second time window.

2. Calculating 1-second the data change range=maximum- minimum.

3. Comparing the 1-second change range with upper control limit (UCL).


|

1A, graph of voltage data CCVT 711 - 1 capacitor fails (short circuit) in phase A at 10 s


|

1C - CCVT 711 - 1 capacitor fails first at 5 s, 2nd capacitor fails after 30sec, in phase A


|

30

• Current State Estimator: Estimate the breaker current. In this model, all the nodesand breakers at the same voltage level inside the substation construct a zero-impedance power system, and the measurement function can be established byapplying KCL. For each branch current, it is a function with respect to two breakercurrents if it is a breaker-and-a-half schema. For each breaker current, it is afunction with respect to itself.

• Voltage State Estimator: Estimate the bus voltage from the voltage measurementsat all the nodes comprising this bus. This is essentially a weighted average and isformulated here as a zero-impedance voltage state estimator. The states are thevoltage of each bus, and the measurements are the voltage phasor measurementsat the nodes belonging to the bus.


|31 31

Voltage Level Separation

500kV 345kV 230kV...

Estimate Current States

Bad Data

Estimate Breaker States

Yes

No

Build the Topology for this

voltage level

Estimate Voltage States

Generate Substation States


Bad Data


No


voltage level



Bad Data


No


voltage level


Abandon

Yes

Abandon

Yes

Abandon




...


|32 32

A substation model of Vassell Substation is created in XML format. The signals of the measurements are also mapped:

Examples of the mapping for different types of measurements are shown below:Voltage:

Breaker Current:


|

1B, graph of voltage data CCVT 711 - 5 capacitor fails (short circuit) in phase A at 10 s

SLSE successfully detected the anomaly caused by CCVT 711 failure


|

1B, CCVT 711 - simulation of temporarily short circuit CCVT secondary at 10 s, (FSC taken out at 0.25 s) clearing around 7 cycles on phase A

SLSE successfully bypassed the anomaly caused by the system fault and did not false alarm for CCVT 711 anomaly



|

Normal operation data without equipment failure nor system event• Each voltage and current signal is tested independently• Didn’t have false alarm based on the setting


|

Validated the accuracy of the SLSE algorithms:• The SLSE didn’t alarm on any anomalies, which is as expected. • The SLSE results are also very close to and following the variations of the raw signals

3 Phase voltage signals:



3 Phase breaker current signals:

|

▪ Working with AEP to establish filed synchrophasor connection to EPG

▪ System integration testing with SLSE and data-driven algorithms

▪ AEP setting up central location VM environment to test on-site

▪ Appreciate if any other utilities can contribute to an equipment failure “data library”

▪ Interested in this project for host demonstration? Still not too late to join!


# SOPO Tasks and Subtasks Planned Timeline

3.0 Development, Testing, and Demonstration July 2017 – August 2018

3.1 Pseudo-Synchrophasor Data July – December 2017

3.2 Field Synchrophasor Data December 2017 – March 2018

3.3 Sampled Data from Instrument Transformers April – August 2018

|

1) NASPI Technical Report, “Diagnosing Equipment Health and Mis-operations with PMU data”,May 2015

2) Bogdan Kasztenny and Ian Stevens, “Monitoring Ageing CCVTs – Practical Solutions withModern Relays to Avoid Catastrophic Failures”, March 2007

3) David Shipp and Thomas Dionise, IEEE Tutorial, “ Switching Transients, Transformer Failures,Practical Solutions”, Feb 2016

4) L. Sevov, J. Cardenas and Y. Sun, "CT Failure Detection For Differential ProtectionApplications," 2008 61st Annual Conference for Protective Relay Engineers, College Station, TX,2008, pp. 498-511. doi:10.1109/CPRE.2008.4515076

5) Deepak Rampersad, “Investigation into current transformer failures within Eskom distribution”,December 2010

6) Darren Spoor and Jian Guo Zhu, Monitoring current transformer secondary circuits to forewarnof catastrophic insulation faults

7) D. Costello, "Open-circuited CT misoperation and investigation," 2014 67th Annual Conferencefor Protective Relay Engineers, College Station, TX, 2014, pp. 383-392,doi:10.1109/CPRE.2014.6799015



[1] Transmission & Distribution Committee - IEEE Power & Energy Society, “Electric Signatures of Power Equipment Failures,” 2015.

[2] M. Al Karim, M. Chenine, K. Zhu, and L. Nordstrom, “Synchrophasor-based data mining for power system fault analysis,” IEEE PES

Innov. Smart Grid Technol. Conf. Eur., pp. 1–8, 2012.

[3] H. Jiang, X. Dai, D. W. Gao, J. J. Zhang, Y. Zhang, and E. Muljadi, “Spatial-Temporal Synchrophasor Data Characterization and Analytics in

Smart Grid Fault Detection, Identification, and Impact Causal Analysis,” IEEE Trans. Smart Grid, vol. 7, no. 5, pp. 2525–2536, 2016.

[4] H. Jiang, J. J. Zhang, W. Gao, and Z. Wu, “Fault detection, identification, and location in smart grid based on data-driven computational

methods,” IEEE Trans. Smart Grid, vol. 5, no. 6, pp. 2947–2956, 2014.

[5] J. M. Lim and C. L. Demarco, “Model-free voltage stability assessments via singular value analysis of PMU data,” Proc. IREP Symp. Bulk

Power Syst. Dyn. Control - IX Optim. Secur. Control Emerg. Power Grid, IREP 2013, 2013.

[6] R. Meier et al., “Power system data management and analysis using synchrophasor data,” 2014 IEEE Conf. Technol. Sustain., pp. 225–

231, 2014.

[7] A. Silverstein, “Diagnosing Equipment Health and Mis-operations with PMU Data,” 2015.

[8] Xiaodong Liang and S. A. Wallace, “Processing synchrophasor data using a feature selection procedure,” in 2016 IEEE PES Asia-Pacific

Power and Energy Engineering Conference (APPEEC), 2016, pp. 273–277.

[9] K. D. Jones, A. Pal, and J. S. Thorp, “Methodology for Performing Synchrophasor Data Conditioning and Validation,” IEEE Trans. Power

Syst., vol. 30, no. 3, pp. 1121–1130, 2015.

[10] N. Dahal, R. L. King, and V. Madani, “Online dimension reduction of synchrophasor data,” Proc. IEEE Power Eng. Soc. Transm. Distrib.

Conf., pp. 1–7, 2012.

[11] J. Ning and W. Gao, “Multi-feature extraction for power system disturbances by wavelet transform and fractal analysis,” IEEE PES Gen.

Meet. PES 2010, pp. 1–7, 2010.

[12] J. Patel, “Real time big data mining,” The State University of New Jersey, 2016.

| 42

Heng (Kevin) Chen

[email protected]

Lin Zhang

[email protected]

Neeraj Nayak

[email protected]



Open Circuit in CT secondary due to Wiring damage

High frequency transients observed 8 minutes before CT failure (partial discharge in insulation)

Normal Operation – No failure Reference: [G]

Reference: [F]


Ferroresonance –

Opening Breaker

Switching

Transients

Loose Connection

at PT feeding the

PMU

Blown fuse on

One Phase of PT

Internal

Primary

Winding Issue

Reference: [A]

Reference: [A]

Reference: [C]


Loose Fuse Connections in CCVT Safety SwitchCapacitor Failure in C phase

A - Phase CCVT Issue

Reference: [A]

Reference: [A]

Reference: [B]

|

Prepare/Smooth Data

Extract Feature

Classify and Quantify Feature

Perform pattern recognition, comparison, etc.

Flag data, record data, execute other

algorithms

Take Action

PMU/LSE Data

No

Yes

1

2

3

4

5

Note: Some algorithms may perform more

than one process in a single step.



Documents

WECC JSIS Meeting€¦ · Post Processing for list of bad measurement Substation Mod el (CIM) Data Record ing Sync h rop as Task 3.2 Pseud o-Synchrophasor Task 3.1 Alarming to Id